Publication details

Exploring trends in quality and features of biomacromolecular complexes with ValTrendsDB



Year of publication 2019
Type Conference abstract
MU Faculty or unit

Central European Institute of Technology

Description One of the most applauded achievements of life sciences is the availability of structural data of biomacromolecular complexes to the scientific public. Unfortunately, the concept of errors in structure models is not just theoretical. The community of structural biologists reacted to this phenomenon by developing and utilizing methodologies and software tools for structure model validation. The emergent focus on validation provoked us to ask: Does it have any impact on quality of new publicly available structures? We have carried out a wide range exploratory analysis of trends in relationships of pairs of factors that represent quality and features of biomacromolecules and ligands. We had two data sources for this analysis: Most data were obtained from the PDB database, while additional information concerning ligand quality was sourced from the ValidatorDB database. Some of the trends that we discovered were expected (e.g., newer structure models have, in general, higher quality than the older ones have), while other discovered trends were a surprise to us (e.g., ligand model quality is stagnant at best). The whole set of results of the exploratory analysis is available in our weekly-updated ValTrendsDB database. Users can view there all factor pair relationship plots of all trends discovered during the analysis. It is also possible to draw custom plots of any pair of factors with custom settings. Additionally, users are able to view value distribution of every factor considered. An important functionality of ValTrendsDB is the support for visualization of data points of a set of interesting PDB entries into any factor pair plot available, regardless of whether such plot is precomputed, or custom. With this functionality, users can view quality and features of their structure models of interest, e.g., structures of a journal, structures of a protein family, structures acquired using a particular experimental method, or structures assembled by specific authors, in relation to the trend across entries of the whole PDB database
Related projects: